Merge branch 'master' of ssh://git-annex.branchable.com into master

This commit is contained in:
Joey Hess 2020-09-28 15:30:23 -04:00
commit e44a132085
No known key found for this signature in database
GPG key ID: DB12DB0FF05F8F38
3 changed files with 54 additions and 0 deletions

View file

@ -0,0 +1,34 @@
I have a repo for which `git annex fsck --all` goes through without any complaints, and `git fsck` produces
Checking object directories: 100% (256/256), done.
error in tree 97abf6de76c0c56593f78d2f55b4c60d04345281: gitmodulesSymlink: .gitmodules is a symbolic link
Checking objects: 100% (998477/998477), done.
Checking connectivity: 998477, done.
dangling tree f9a2b542f57770c135631de1ae590475fc0caa54
but running `git annex repair --force`,
repair Running git fsck ...
Unpacking all pack files.
Unpacking objects: 100% (998477/998477), 64.95 MiB | 359.00 KiB/s, done.
Initialized empty Git repository in /tmp/tmprepoqPmkhi/.git/
Trying to recover missing objects from remote A.
fatal: '[a path I censored]' does not appear to be a git repository
fatal: Could not read from remote repository.
Please make sure you have the correct access rights
and the repository exists.
Trying to recover missing objects from remote B.
Unpacking all pack files.
Unpacking objects: 100% (923941/923941), 63.54 MiB | 367.00 KiB/s, done.
[A couple of other cases of remotes not available.]
failed
git-annex: repair: 1 failed
Is the failure likely related to the `gitmodulesSymlink` error? If so, what could I do about it? If not, how do I figure out what's going on?
Regarding the `.gitmodules` complaint: The repo has a subfolder with its own `.git` and `.gitmodules`, and although nothing in there is commited into the main repo now, I did in the past accidentally, temporarily commit the `subfolder/.gitmodules` into the main repo.
I don't actually find anything wrong with the repo, the only reason I'm running `repair` on it is because a remote of it is missing some files, although it claims to be all synced up.
This is on version 8.20200226. I would be happy to try on a more recent one, if there's a way to do so on Ubuntu without compiling from source.

View file

@ -0,0 +1,8 @@
[[!comment format=mdwn
username="mhauru"
avatar="http://cdn.libravatar.org/avatar/2532433a0207ba772e6ca964e61899c0"
subject="NeuroDebian seems to have stopped updating"
date="2020-09-27T12:40:54Z"
content="""
The comment about NeuroDebian providing more up to date builds is out of date, their version is now more than a year old. An alternative would be welcome, something between the official Ubuntu package and compiling from source.
"""]]

View file

@ -0,0 +1,12 @@
[[!comment format=mdwn
username="yarikoptic"
avatar="http://cdn.libravatar.org/avatar/f11e9c84cb18d26a1748c33b48c924b4"
subject="comment 16"
date="2020-09-27T13:42:15Z"
content="""
Eh, true. Our version is year old as the one which doesn't cause regressions in datalad. We are still fighting with some remaining regressions. See eg one of the last attempts https://github.com/datalad/datalad/pull/4915 .
We do have unannounced on the NeuroDebian debian-devel/ (complement to regular debian/ line in your apt listing) from the main NeuroDebian website (not mirrors). That is where we upload those versions to test.
Also there is now a daily built git annex .deb (and .dmg for osx) package as artifact of GitHub workflow, eg https://github.com/datalad/datalad-extensions/actions/runs/274613689 . I guess we better automatically upload them somewhere for easier fetching/deployment.
"""]]